Leaking of questions: BCS examination

We wrote an article in The Financial Express on May 13 illustrating through an economic analysis how centralised control of the medical entrance examination creates higher possibility of questions leaking. In addition, we explained how the centralised method of distribution of students among different medical colleges creates the inefficient allocation problem. We advanced the argument that decentralisation of the medical examination procedure - giving each medical college the authority to make and conduct examination with unique question sets - would reduce such leaking possibility while ensuring an efficient allocation of students among different medical schools.
Can such an argument can be advanced in the case of BCS (Bangladesh Civil Service) examination? Will the decentralisation technique reduce the leaking possibility of BCS question papers? We say that decentralised system will not work properly to prevent leaking of BCS question papers though it will work effectively in the case of medical entrance examination. In this article we will illustrate a mechanism that can prevent leaking of BCS question papers while demonstrating why the decentralisation technique will not work for the BCS examination.
WHY WILL DECENTRALISATION METHOD NOT WORK? For the medical entrance examination, we calculated the probability of leaking question both under the centralised and the decentralised method of conducting examination. We showed that probability of leaking question under the decentralised system is almost near to zero while the probability of leaking under the centralised system is close to one. The calculated probability of leaking question under the centralised system was given by [p(L)=p(m)p(q){f(ck-1)}] where p(L) is the probability of leaking question, p(m) probability of leaking question through selecting medical schools, p(q) probability of leaking question if question sets are not homogenous and f(ck-1) probability of leaking question at a given information collection cost. The principal assumption on which this probability was calculated was that individuals follow some optimising rules and they have sufficient incentive to follow these rules. If the incentives to follow the optimising rules under the decentralised system of examination are absent or are made weak, then this probability measure will not the match real world's probability of leaking examination question.
This is the tragedy that plagues the BCS examination system. The peculiarity of this system is that decentralisation will not reduce the probability of leaking examination question unless some controlled factors are made ineffective. To present the analysis in more vivid way, we present the following graphical illustration.
In the above figure, probability of leaking question p(L) is measured on the vertical axis and information collection cost(c) on the horizontal axis. The graph shows that as the information collection cost increases, probability of leaking question decreases and vice-versa. The final probability given at the cost (ck-1) under the decentralised system is p(m)p(q)[f(ck-1) which is close to zero.  Let's say that under the current centralised system of BCS examination, conducted and managed by the Public Service Commission (PSC), probability of leaking of BCS question papers is p0 which is very high and close to one. From the graph, we see that probability of leaking under decentralised exam system is p(m)p(q)[f(ck-1)]. But we see that probability of leaking is still p0 even at the information cost c0k-1. But at c0k-1 the probability of leaking should be p(m)p(q)[f(ck-1)]. Therefore, there is no systematic relationship between probability of leaking BCS question papers and the decentralisation of examination system.
Why is the decentralisation technique effective in reducing the probability of leaking of medical entrance examination question but is ineffective in reducing probability of leaking BCS examination questions?
CONTROLLED FACTORS AND LEAKING POSSIBILTY: To understand why decentralisation method will not reduce the possibility of leaking of BCS examination questions, we need to understand what kind of factors cause leaking of questions. We mentioned above whether our calculated probability of leaking of questions under the decentralised method will match the real world's probability of leaking depends on whether people follow the optimising rules as to decide which medical school to select and with whom to compete taking information collection cost. We distinguish between two effects: Manipulation effect and stimulus effect. First one increases the leaking possibility and the second one dampens the possibility of leaking.  We need to know what causes the manipulation effect and how this effect counteracts and dominates the stimulus effect generated by the decentralisation method to generate higher leaking possibility effect.
To understand this, we can visualise the following scenario. Let's say that if the examination system is decentralised, it will give people incentive to follow optimising rules since leaking of examination questions is not a worthy attempt given the toil and effort it will require. We designate such rules as R. if people follow these rules, then we will see an outcome which we designate as V(R) that depends on R. This outcome is the probability of leaking. Our objective is to assess the following condition: whether the real probability of BCS questions leaking termed as V(Z) match our calculated probability of BCS questions leaking p(m)p(q)[f(ck-1)] or [V(Z) = p(m)p(q){f(ck-1)}]  if the system is decentralised. Now, say that there are some factors retained and controlled by centralised authority, PSC that conducts the BCS examination. We lump these factors under Z that determines how strong the manipulation effect and how much higher the leaking possibility is. The greater is the manipulation power, the higher manipulation effect and leaking possibility. Then we quantify the manipulation effect by quantifying X which measures the level of corruption. Now whether people have sufficient incentives to follow rule R or not depends on whether the value of X is higher or smaller which depends ultimately on Z. Concisely, we say as follows:

 Now we can comfortably measure the comparative effect as to which one dominates: manipulation effect or stimulus effect. The manipulation effect is measured by V(Z) and stimulus effect by V(R).  So, we say:

THE AXE IN THE WRONG HAND: In the above box we have shown why the actual probability of leaking of BCS examination question and the calculated probability of leaking don't match and why the probability of leaking remains independent of the system of examination - whether centralised or decentralised. Now we will explain the reason in plain language. We have described that there are some factors that are retained by centralised authority, PSC and that determines the manipulation power of the authority. This manipulation power determines the level of corruption as designated by X in the above box. The manipulation power gives incentives to people not to follow the optimising rules generated by the decentralisation of the examination system. Under the weak manipulation power, people get sufficient incentives to follow the optimising rules that induce them not to hanker after leaking of question papers. But as long as manipulation power is retained by PSC and its allied forces such as political bodies, people get enough incentive to hanker after leaking of question papers and most often turn out successful because of the existence of the manipulation power. Moreover, our political practices are too corrupt to be transparent and effective. In most of the cases, manipulation power is directly and indirectly created and supported by political liaisons. BCS examination is such an examination that suffers from unethical political practices. And for this reasons, people get strong motivation to look for any opportunity to leak question papers. The real culprit is the manipulation power concentrated in the hand of the centralised authority that prepares and manages question papers. Therefore, it seems that entrusting the duty of preparing and managing question papers of BCS examination on PSC means putting axe in the wrong hand. So, the question arises: what is the way out from this question papers leaking problem?
THE SOLUTION: The solution is quite straightforward. We suggest that the government should take away the responsibility of preparing and managing question papers from the PSC and try to devise another way for preparing and managing BCS question papers. One feasible way that can ensure better transparency without generating higher leaking possibility is outsourcing. Instead of entrusting the responsibility of preparing and managing question papers to the PSC that contains the higher possibility of leaking, the government, if it wants better transparency and efficient use of resources, should outsource question papers. Outsourcing method that is quite free from unwanted manipulation power created by both direct and indirect liaison with the ruling government and the improper use of power can render much better solution. The benefit would be two fold. First, outsourcing question papers would reduce leaking possibility of BCS question papers. The reason is when government would outsource BCS question papers, it would demand more transparency since the body making and preparing question papers would not totally be supported and controlled by the government and its allied forces. Moreover, outsourcing means curbing corruption. Since the current centralised authority belongs to government it would show slackness in controlling corruption as like minds rarely spoil like minds. But outsourcing means binding two unlike minds and this binding would generate more transparency and accountability from both sides. Second, the government would be able to reduce its inefficient use of resources reducing its protracted and inefficient management of BCS examination through the centralised control. The outsourcing would reduce the size of centralised control which will reduce the use of funds in inefficient uses. The outsourcing would enable government to minimise its burdensome control and keep the size of control to the minimum inefficient level and the maximum efficient level.
We are not, however, arguing that the alternative way would be totally free of manipulation power. But we are arguing for comparative transparency and the outsourcing may ensure that.
The writer is with the University
of Denver, USA.
 [email protected]